An automated brain tumor segmentation and classification using adaptive Bayesian fuzzy clustering

An uncontrolled growth of malignant cells in the brain is known as a brain tumor. Rapid treatment response follows an early identification of tumors in the brain that increases the chance of patient survival. Adequate tumor classification and segmentation are necessary for treatment planning and best evaluation. It would be ideal and beneficial to have regular detection and identification. The design of medical imaging systems has been greatly influenced by the introduction of deep learning in recent years. Hence, an innovative brain tumor classification model is suggested in this work that resolves the drawbacks of traditional methods such as computational complexity, and low accuracy. At first, the necessary images are garnered from the online benchmark for the subsequent process. Further, the garnered images are given to the segmentation procedure, where an Adaptive Bayesian Fuzzy Clustering (ABFC) is utilized for segmenting the abnormalities. Moreover, an Improved Eurasian Oystercatcher Optimizer (IEOO) is adopted in the segmentation process for tuning the parameters in the ABFC technique, which increases the performance. The segmented images are subjected to the Multi-scale Residual Attention Network with Long Short Term Memory (MRAN-LSTM) layer for classifying the brain tumors. Finally, the simulations are done to verify the success rate of the implemented brain tumor segmentation and classification approach by contrasting it with traditional models.